Research Onautomatic Recognition Of Theanalog Instrument Based On Image Processingtechnology

In some fields of application, people need to know the insulation of resistances. Analog instrument is widely applied in actual projects. The reading of it generally depends on manual observation operation or AD sampling digitization reading. This is insufficient in precision, and wastes both time and energy. Using computer vision and image processing technology, automatic reading of analog instrument can be realized, which has much higher efficiency, less consumption and lower production costs. In order to realize automatic reading, it is necessary to take processing on the meter display image and to identify the pointer information.This thesis first introduces the overall framework of the automatic reading system of analog instrument, and also analyzes each component unit. Then researches on reading methods of the recognition, which is the focal point of this thesis. Starting from the image processing process, the thesis carries on the introduction and the research to each step, and also gives out the effect which each step processes. Through the image difference method, the pointer is extracted fast and easily, and the extraction is very effective. In scale division distinction aspect, scale mapping method is adopted, which is the way of mapping circular arc to straight line. The value of the pointer is recognized out correctly by polynomial fitting method. In the error analysis, this article gives a detailed way. Considered the vibration of the image photography, reading accuracy will be affected, so the thesis conducts the research to the image registration method, which can reduce errors. The user application is given out in the end.